We have prepared single-chain immunoglobulin Fv fragments from the CD20-specific hybridoma HB13d. One scFv clone demonstrated strong binding to a CD20-derived peptide by ELISA and to CD20-positive cells by flow cytometry, a second had reduced binding, and a third clone did not bind the target antigen. Sequence analysis showed that all three constructs contained shared and unique amino acid changes when compared to the nearest germline match. Molecular modelling of the scFv variants revealed that several of the mutations are located in regions predicted to contact antigen, including a mutation in the heavy chain CDR1 of the strongest binding scFv construct. No similar mutation is present in the highly conserved protein sequences of a number of CD20-specific monoclonal antibodies. BIACORE analysis demonstrated that the mutated scFv had approximately three-fold greater antigen-binding activity than another clone. Competition studies showed that the scFv is able to compete with intact CD20 monoclonal antibody for binding to the target antigen. The improved antigen binding of this scFv will permit the construction of novel CD20-specific reagents for the therapy of lymphomas.
The CD20 antigen is a 35 kDa phosphoprotein expressed on the surface of B lymphocytes from the pre-B stage until terminal differentiation to plasma cells. The amino acid sequence of the CD20 antigen predicts four transmembrane spanning regions, with both amino and carboxyl terminals located on the cytoplasmic side of the plasma membrane (Tedder et al., 1989). CD20 antigen is a member of a multigene four-transmembrane family (termed MS4A; Liang and Tedder, 2001), which is distinct from the tetraspanin superfamily that includes the B-cell antigens CD9 and CD37 (Kitadokoro et al., 2001). The function of the CD20 antigen is not fully understood, although it is believed to have a role as a calcium-ion channel that is associated with detergent-resistant lipid raft microdomains in the cell membrane (Bubien et al., 1993, Li et al., 2004). Disruption of the CD20 gene has a minimal effect on B-cell differentiation or function in mice (O’Keefe et al., 1998, Uchida et al., 2004).
CD20 is expressed by more than 95% of normal B cells isolated from peripheral blood, lymphoid tissues, and bone marrow (Stashenko et al., 1980). The antigen is also expressed on the malignant cell population in 50% of non-T-cell acute lymphoblastic leukemias (ALL), greater than 95% of B-cell chronic lymphocytic leukemias (CLL) and greater than 90% of non-Hodgkin's B-cell lymphomas (NHL). Significant therapeutic benefit has been obtained by treating low-grade NHL with CD20 monoclonal antibodies such as Rituximab, a genetically engineered chimeric CD20 antibody (Coiffier et al., 1998, Reff et al., 1994), with an overall response rate of 48% in patients with relapsed or refractory low-grade or follicular NHL (reviewed in White et al., 2001). The CD20 antigen has advantages as a therapeutic target over other B lineage markers, such as CD19 and CD22 because it is expressed at higher levels and is not shed or internalised after cross-linking by antibody (Press et al., 1989).
Genetically engineered antibody fragments offer a number of potential advantages as therapeutic agents over whole antibodies (for a review see Hudson and Souriau, 2003). Using hybridomas as the starting material, single-chain Fv fragments (scFv) which retain the specificity and many of the antigen binding characteristics of the parent can be constructed as a genetic fusion of the heavy chain and light chain antigen-binding domains, linked by a flexible polypeptide (Clackson et al., 1991, Krebber et al., 1997). The ability to readily engineer scFv-based fusion proteins has seen CD20-specific scFv used to produce therapeutic agents with a variety of effector functions (Haisma et al., 1998, Jensen et al., 1998, Schultz et al., 2000, Wu et al., 2001).
Here, we describe the construction and functional characterisation of three CD20-specific scFv derived from the HB13d hybridoma, one of which demonstrated superior antigen binding. Sequence analysis and molecular modelling of the three scFv variants suggests that the differences in antigen binding may be due to the mutations in the heavy-chain CDR1. The superior scFv was able to inhibit antigen binding by an intact CD20 monoclonal antibody, indicating that it may be useful for the preparation of novel CD20-specific therapeutic agents.
A CD20-specific single-chain Fv construct was prepared from HB13d hybridoma mRNA (Kansas and Tedder, 1991) using the primer sets and pAK series of vectors described by Krebber et al. (1997), which were provided by Dr. Andreas Plückthun (Biochemisches Institut der Universitat Zurich, Switzerland). PCR amplification, scFv assembly and cloning, and the selection of expressing colonies were performed as described previously (Mavrangelos et al., 2001).
The pHB400 vector (Mavrangelos et al., 2001)
We identified three c-myc tag-expressing clones following transformation of the HB13d-derived scFv construct. Two of these clones, HB13d-7 and HB13d-11, expressed a functional scFv protein that bound to Raji cells as shown by flow cytometry (Fig. 1A), while a third, HB13d-4, expressed a product that did not bind Raji cells. ELISA analysis confirmed that the HB13d-7 and HB13d-11 scFv were reactive with a peptide derived from the CD20 antigen, but that the HB13d-4 scFv was not active (Fig. 1B).
The CD20 antigen is currently used as a target for antibody-mediated therapy of B-lineage lymphomas and leukemias (Coiffier et al., 1998). More recently, antibodies to CD20 have been used to treat as a number of other conditions that involve B cells (Rastetter et al., 2004). Cross-linking of the CD20 monoclonal antibodies B1, 2H7 and 1F5 induces calcium-dependent apoptosis in the Ramos B cell line (Deans et al., 1998, Shan et al., 1998). Segregation of the antigen into lipid rafts and
We thank Silvia Nobbs and Sheree Bailey for assistance with flow cytometry. These studies were supported by an Australian Postgraduate Award (Industry) to PJA, a Project Grant from the Australian National Health and Medical Research Council, and a grant-in-aid from the Childhood Cancer Association, Adelaide.
